5-whys

# 5 Whys ## Overview The 5 Whys is an iterative interrogative technique developed by Sakichi Toyoda in the 1930s and formalized by Taiichi Ohno at Toyota Motor Corporation as a cornerstone of the Toyota Production System. The method explores cause-and-effect relationships by repeatedly asking "why?" (typically five times) until the root cause emerges. The number five is not rigid - continue until you reach a systemic root cause rather than a symptom. Ohno called it "the basis of Toyota's scientific approach" to problem-solving. The technique is deceptively simple but powerful: each answer becomes the basis for the next question, creating a chain of causation that reveals where intervention will have lasting impact rather than temporary fixes. ## When to Use - A problem keeps recurring despite multiple "fixes" (symptom treatment, not root cause) - You need quick root cause analysis without complex statistical tools - Simple to moderately complex problems with human/process factors - Team needs shared understanding of why a failure occurred - Manufacturing defects, process breakdowns, or organizational issues - You suspect the obvious answer masks a deeper systemic issue ## The Process ### Step 1: Define the Problem Precisely State the problem as a specific, observable fact. Avoid vague descriptions. **Example:** "The machine stopped working at 2:15 PM on Line 3" (not "machines are unreliable") ### Step 2: Ask "Why Did This Happen?" - First Why Answer based on facts, not speculation. Gather data from people who witnessed the problem. **Example:** Why did the machine stop? → The motor overheated and the thermal fuse blew. ### Step 3: Ask "Why?" of the Previous Answer - Second Why Take the first answer and ask why that condition existed. **Example:** Why did the motor overheat? → The bearing was not sufficiently lubricated. ### Step 4: Continue the Chain - Third, Fourth, Fifth Whys Keep drilling down. Each answer should point to a more fundamental cause. **Example (Ohno's classic):** - Why wasn't it lubricated? → The lubrication pump wasn't pumping sufficiently. - Why wasn't it pumping? → The pump shaft was worn and rattling. - Why was the shaft worn? → No strainer was attached; metal scraps got in. ### Step 5: Identify the Root Cause and Implement Systemic Fix The final "why" reveals the root cause. Fix this, not the symptoms. **Root cause:** No strainer on pump intake **Fix:** Install strainer + add inspection checklist for all pumps ### Step 6: Verify the Fix Prevents Recurrence Test that addressing the root cause eliminates the problem completely. **Verification:** Run machine for 100 hours with new strainer - no lubrication failures. ## Example Application **Situation (Taiichi Ohno, Toyota):** Machine stopped on production line. **Application:** 1. Why stopped? → Overload blew the fuse 2. Why overload? → Bearing not lubricated 3. Why not lubricated? → Pump not pumping sufficiently 4. Why not pumping? → Pump shaft worn and rattling 5. Why shaft worn? → No strainer; metal scraps got in **Outcome:** Root cause identified as missing strainer. Installing strainers across all pumps prevented recurrence. Without 5 Whys, they would have replaced the fuse (symptom) and the problem would repeat. ## Anti-Patterns - ❌ Stopping at symptom-level answers ("the fuse blew" - replace fuse without asking why) - ❌ Accepting opinions instead of facts ("someone was careless" vs. "no checklist exists") - ❌ Using 5 Whys for highly complex problems requiring statistical analysis - ❌ Rigidly asking exactly five questions when root cause appears at question 3 or requires 7 - ❌ Working alone instead of gathering the team closest to the problem - ❌ Jumping to solutions before completing the causal chain - ❌ Blaming people rather than identifying systemic/process failures ## Related - fishbone-diagram (visual root cause analysis with categorized causes) - fault-tree-analysis (deductive failure analysis for complex systems) - pre-mortem (imagining failure before it happens) - inversion (identifying what would cause failure) - second-order-thinking (tracing consequences beyond first level)